The SupaDevs is a searchable library of developer tools, UI components, APIs, tutorials, and other resources.
Why I built it
I kept finding useful developer resources and then losing them in bookmarks, chats, and browser tabs. The SupaDevs started as my attempt to put those resources in one searchable place. I also wanted to make contributing easier, so I explored using AI to categorize submissions instead of making every contributor organize everything manually.
| Year | 2025 |
| Type | Web app |
| Client | Personal project |
| My role | Full-stack development, AI integration |
| Platform | Web |
| Link | The SupaDevs |
The platform combines resource discovery with metadata extraction and AI-assisted categorization, helping submissions become easier to browse and manage.
200+
resources listed
5 min
configured ISR revalidation
AI
resource categorization
Challenge
Resources come from different websites, each with its own metadata conventions. The challenge was extracting useful information consistently, organizing submissions, and giving administrators a practical way to review them.
Action
Metadata extraction and fallbacks
I built a scraper that checks Open Graph metadata first, then Twitter card data, page images, and finally the favicon. Caching, retries, and rate-limit handling help manage requests to third-party sites.
AI-assisted categorization
I integrated Groq with Llama 3.1 8B to categorize batches of submitted resources and suggest tags. The workflow also produces confidence scores to help identify classifications that may need review.
Search, caching, and admin tools
I built the site with Next.js and ISR, using five-minute revalidation for the configured pages. Admin tools support bulk approval, link checks, and resource categorization.
Result
The SupaDevs combines resource discovery with automated metadata extraction and AI-assisted organization. It gave me practical experience with caching, web scraping, and integrating language models into an existing workflow.